カオスマップに戻る

Overview

Claude is Anthropic's conversational AI app, available on Web, iOS, Android, and desktop. It is favored by developers and knowledge workers for its strengths in coding, long-context reasoning, and safety. As of March 2026, Claude runs on the Opus 4.6 / Sonnet 4.6 flagship models with 1M token context at standard pricing.

Latest / Key Features

Claude 4.6 Series

Since March 2026, Pro and higher plans use Opus 4.6 and Sonnet 4.6. The 1M token context window is now GA at standard rates, so large codebases and long documents carry no premium cost.

Cowork (Agentic Feature)

Launched for Max on January 12, 2026, and extended to Pro four days later. Cowork lets Claude autonomously complete multi-step tasks such as organizing research files, analyzing data across multiple documents, and coordinating project workflows.

Pay-as-You-Go Overages

A major 2026 change: Pro, Max 5x, and Max 20x all now support continuing past plan limits at API-standard pricing, with user-set monthly caps.

Artifacts / Projects

Artifacts renders code, docs, and mini apps in a live preview during chat. Projects persists long-running context and files. Both are available on Pro and above.

Claude Code Integration

Max plans ($100 / $200) include full access to Claude Code, the terminal-based autonomous coding agent that understands entire repositories, edits across files, and runs commands on its own.

Plans & Pricing

PlanMonthlyHighlights
Free$0Sonnet 4.6 with limits
Pro$20Opus 4.6, Projects, Artifacts, Cowork, overages
Max (5x)$1005x Pro usage, full Claude Code
Max (20x)$20020x Pro usage, priority access
TeamFrom $25/userTeam features, admin console
EnterpriseContact salesSSO, audit logs, extended context, SLA

Common Use Cases

  • Software development: Claude Code, Artifacts, repository-level review
  • Long-document work: Contracts, papers, books (1M tokens)
  • Writing: Blogs, technical docs, marketing
  • Data analysis: Cross-document analysis automated via Cowork
  • Research: Deep reasoning, nuanced critique, summarization

Strengths

  • Industry-leading coding ability — tops SWE-bench leaderboards
  • 1M token context at standard pricing
  • Cowork agent — available even on Pro
  • Safety-first — Constitutional AI approach
  • Flexible overages — pay-as-you-go beyond plan limits
  • Claude Code included on Max plans

Weaknesses

  • ❌ No image or video generation (input only)
  • ❌ No app store equivalent to Custom GPTs
  • ❌ Limited voice mode
  • ❌ Tight Free plan limits
  • ❌ Fewer third-party plugins than ChatGPT

Official Links

公式サイト